Oversampling factor to reach a target subgroup composition

The multiplier applied to the subgroup's selection probability so that it forms a stated share of the achieved sample rather than its natural share of the frame. A factor above 1 means disproportionate stratified selection, which buys subgroup precision at the cost of unequal weights; the resulting weight variation feeds back into the design effect, so the two should be planned together rather than in sequence.

Usage

morie_oversample_factor(domain_prevalence, target_share)

Arguments

domain_prevalence

Share of the frame belonging to the subgroup.

target_share

Desired share of the achieved sample.

Value

A list with factor (the selection-probability multiplier), weight_ratio (the ratio of subgroup to non-subgroup design weights that results), and deff_weights, the Kish design effect implied by those two weights at the target composition.

See also

morie_sample_size_domain, morie_design_effect

Examples

# Lift a 5% subgroup to 20% of the achieved sample
morie_oversample_factor(0.05, 0.20)
#> $factor
#> [1] 4.75
#> 
#> $weight_ratio
#> [1] 0.2105263
#> 
#> $deff_weights
#> [1] 1.140625
#>